1. The Field of the Invention
The present invention is directed generally to apparatus for improving laser technology. More specifically, the present invention is related to apparatus for improving the quality of semiconductor diode lasers, broad area diode lasers, diode laser bars, and diode laser arrays.
2. The Relevant Technology
A laser is a device which utilizes the transitions between energy levels of atoms or molecules to amplify or generate light. When an electron makes a transition from a higher energy level to a lower energy level, a photon, the elementary quantity of radiant energy, is emitted. In what is referred to as "stimulated emission," an incoming photon stimulates an electron to change energy levels, which amplifies the number of exiting photons. In fact, this is the origin of the term laser: light amplification by the stimulated emission of radiation. The emitted photon travels in the same direction and is in the same phase as the incoming photon. When the stimulated emission in a laser involves only a single pair of energy levels, the resultant output beam has a single frequency or wavelength and is thus approximately monochromatic.
One example of a conventional laser is the semiconductor laser diode. Semiconductor lasers are particularly useful for several reasons: they are capable of generating coherent radiation in the wavelength range which is particularly useful for optical fiber communications; they are relatively easy to fabricate and less costly than conventional, larger gas lasers; and they have a compact size which is useful in many applications including optical fiber communications, printing, and medical treatments.
One example of a conventional semiconductor laser is a broad area laser diode. The "broad area" refers to the junction plane from whence the laser radiation originates. Most broad area semiconductor lasers comprise a "stripe" geometry. The stripe geometry typically has dimensions of about 5 to 30 .mu.m in width, about 0.1 mm in length, and about 1 .mu.m in thickness. There are several advantages with respect to this geometry. First, there is improved response time due to small junction capacitance. Further, the thin active layer which is the area wherein the laser radiation is generated and confined, contributes to a smaller cross sectional area. This reduces the operating current, which is necessary for sustained operation of the laser, and also reduces the threshold current, which is the current required to induce a laser device to commence lasing action.
However, the laws of diffraction dictate that beam divergence, which is not desirable, will greatly increase with decreasing aperture size. Yet, most applications require a small beam with maximum power in the smallest area possible. Increasing the width of the aperture does not help, because not only does it reduce the power per area of the emitted beam, it has been demonstrated that the modal characteristics are significantly degraded as the aperture width is increased past a certain point. As this width is increased, the mode degrades from a single, good quality Guassian intensity profile, to several filaments of the beam spread over a large area.
Another example of a semiconductor laser is a laser diode bar. Attempts have been made to increase the power with respect to these semiconductor lasers by combining multiple laser diodes into what is termed a "laser array." The advantage of placing individual laser diodes into an array is that the overall output power can be increased by phase locking several diode lasers together such that they operate as a single source. Yet, even though the power does increase when multiple lasers are combined to produce multiple beams, the quality is extremely poor. In turn, users are forced to spend increased time and money in attempting to alleviate the poor laser quality of the laser arrays, with less than ideal results.
In addition, achieving the desired mode control and coherence from arrays has proved difficult. As a result, virtually all the high-power arrays commercially available emit their radiation into two broad far-field lobes instead of a single diffraction-limited lobe.
Over the last decade there has been a tremendous amount of research effort spent in designing and fabricating high power arrays with adequate modal control and degree of coherence. Commercially available diode laser arrays have been available for the last few years which utilize stacked configurations of bars of laser diodes which lie in the grooves of a planar substrate containing a heat sink for the device. These stacked diode bars use a technology which is built upon "rack and stack" configurations. See, e.g., U.S. Pat. No. 5,311,535 and U.S. Pat. No. 5,526,373 by Karpinski, both of which are incorporated herein by reference.
Yet, the use of diode laser bars in this stacked design has many disadvantages. For example, this stacked design is inflexible and limited to a planar configuration. In addition, each diode laser bar, typically including more than twenty individual laser diodes, is pumped as an integral unit and individual laser diodes cannot be pumped or replaced separately. If even one diode laser inside the laser bar is damaged, the entire bar must be replaced. Unfortunately, replacement of one laser diode array containing only one diode laser bar is expensive.
Furthermore, the emitted laser beams from laser diode arrays experience significant divergence. This problem is addressed in U.S. Pat. No. 5,311,535 and U.S. Pat. No. 5,668,825 by Karpinski, the disclosures of which are incorporated herein by reference. Specifically, individual lenses are placed at a predetermined distance with respect to each diode laser. The radiation emitted from each diode laser passes through a lens which collimates the laser beam. Such a system requires the fabrication of multiple microlenses and the accurate placement of each, which complicates the manufacturing process thereby raising the overall cost of the system. Furthermore, any misalignments in the placement of the lenses greatly reduces the efficiency of the system, yet adjustments in alignment are extremely difficult. This system at its best is able to convert electrical power into optical power at an efficiency of about 50%.
U.S. Pat. No. 5,333,077, by Legar, the disclosure of which is incorporated herein by reference, suggests an alternative solution to the problem of the diverging emission which comprises a combination of aperture filling, and geometrical transformation, and requires two optical elements. The first optical element is a linear array of lenses, each of which receives and redirects a different portion of the emission from the diode array. The second optical element is a separate two-dimensional array of lenses located at the imaginary plane and aligned with the two-dimensional pattern of light beams thereby redirecting and focusing the two dimensional pattern of light beams to the focal point. The second array of lenses can be placed in a rectangular or a hexagonal close packed geometrical pattern. The lenses in both arrays are diffractive lenses.
Although the efficiency could in theory approach approximately 99%, the fabrication of these diffractive lens arrays with the appropriate profiles is difficult, time consuming, and costly. This is due in particular to the complex mask and etch technique commonly used in integrated circuit fabrication. Masks must be made and etching must be done at each step. For a lens of m levels, there must be P master masks made and P etching steps performed, where m=2.sup.P. In addition to the fabrication constraints, the alignment of the microlens arrays must be very precise with respect to each other and with respect to the diode array.